The name Austroasiatic was coined by Wilhelm Schmidt (German: austroasiatisch) based on auster, the Latin word for "South" (but idiosyncratically used by Schmidt to refer to the southeast), and "Asia".^[7] Despite the literal meaning of its name, only three Austroasiatic branches are actually spoken in South Asia: Khasic, Munda, and Nicobarese.

Regarding word structure, Austroasiatic languages are well known for having an iambic "sesquisyllabic" pattern, with basic nouns and verbs consisting of an initial, unstressed, reduced minor syllable followed by a stressed, full syllable.^[8] This reduction of presyllables has led to a variety of phonological shapes of the same original Proto-Austroasiatic prefixes, such as the causative prefix, ranging from CVC syllables to consonant clusters to single consonants among the modern languages.^[9] As for word formation, most Austroasiatic languages have a variety of derivational prefixes, and many have infixes, but suffixes are almost completely non-existent in most branches except Munda, and a few specialized exceptions in other Austroasiatic branches.^[10]

The Austroasiatic languages are further characterized as having unusually large vowel inventories and employing some sort of pitch register contrast, either between modal (normal) voice and breathy (lax) voice or between modal voice and creaky voice.^[11] Languages in the Pearic branch and some in the Vietic branch can have a three- or even four-way voicing contrast.

However, some Austroasiatic languages have lost the register contrast by evolving more diphthongs or in a few cases, such as Vietnamese, tonogenesis. Vietnamese has been so heavily influenced by Chinese that its original Austroasiatic phonological quality is obscured and now resembles that of South Chinese languages, whereas Khmer, which had more influence from Sanskrit, has retained a more typically Austroasiatic structure.

Much work has been done on the reconstruction of Proto-Mon–Khmer in Harry L. Shorto's Mon–Khmer Comparative Dictionary. Little work has been done on the Munda languages, which are poorly documented. Proto-Mon–Khmer becomes synonymous with the Proto-Austroasiatic language with their demotion from a primary branch. Paul Sidwell (2005) reconstructs the consonant inventory of Proto-Mon–Khmer as follows:^[12]

This is identical to earlier reconstructions except for *ʄ. *ʄ is better preserved in the Katuic languages, which Sidwell has specialized in.

Linguists traditionally recognize two primary divisions of Austroasiatic: the Mon–Khmer languages of Southeast Asia, Northeast India, and the Nicobar Islands, and the Munda languages of East and Central India and parts of Bangladesh and Nepal. However, no evidence for this classification has ever been published.

Each family written in boldface below is accepted as a valid clade.^{[clarification needed]} By contrast, the relationships between these families within Austroasiatic are debated. In addition to the traditional classification, two recent proposals are given, neither of which accepts traditional "Mon–Khmer" as a valid unit. However, little of the data used for competing classifications has ever been published and, therefore, cannot be evaluated by peer review.

In addition, there are suggestions that additional branches of Austroasiatic might be preserved in substrata of Acehnese in Sumatra (Diffloth), the Chamic languages of Vietnam, and the Land Dayak languages of Borneo (Adelaar 1995).^[13]

Diffloth's widely cited original classification, now abandoned by Diffloth himself, is used in Encyclopædia Britannica and—except for the breakup of Southern Mon–Khmer—in Ethnologue.

Peiros is a lexicostatistic classification, based on percentages of shared vocabulary. This means that languages can appear to be more distantly related than they actually are due to language contact. Indeed, when Sidwell (2009) replicated Peiros's study with languages known well enough to account for loans, he did not find the internal (branching) structure below.

Diffloth compares reconstructions of various clades, and attempts to classify them based on shared innovations, though like other classifications the evidence has not been published. As a schematic, we have:

Or in more detail,

Paul Sidwell (2009), in a lexicostatistical comparison of 36 languages which are well known enough to exclude loanwords, finds little evidence for internal branching, though he did find an area of increased contact between the Bahnaric and Katuic languages, such that languages of all branches apart from the geographically distant Munda and Nicobarese show greater similarity to Bahnaric and Katuic the closer they are to those branches, without any noticeable innovations common to Bahnaric and Katuic.

He therefore takes the conservative view that the thirteen branches of Austroasiatic should be treated as equidistant on current evidence. Sidwell & Blench (2011) discuss this proposal in more detail, and note that there is good evidence for a Khasi–Palaungic node, which could also possibly be closely related to Khmuic.^[15]

If this would the case, Sidwell & Blench suggest that Khasic may have been an early offshoot of Palaungic that had spread westward. Sidwell & Blench (2011) suggest Shompen as an additional branch, and believe that a Vieto-Katuic connection is worth investigating. In general, however, the family is thought to have diversified too quickly for a deeply nested structure to have developed, since Proto-Austroasiatic speakers are believed by Sidwell to have radiated out from the central Mekong river valley relatively quickly.

Subsequently, Sidwell (2015a: 179)^[16] proposed that Nicobarese subgroups with Aslian, just as how Khasian and Palaungic subgroup with each other.

A subsequent computational phylogenetic analysis (Sidwell 2015b)^[17] suggests that Austroasiatic branches may have a loosely nested structure rather than a completely rake-like structure, with an east–west division (consisting of Munda, Khasic, Palaungic, and Khmuic forming a western group as opposed to all of the other branches) occurring possibly as early as 7,000 years before present. However, he still considers the subbranching dubious.

Integrating computational phylogenetic linguistics with recent archaeological findings, Paul Sidwell (2015c)^[18] further expanded his Mekong riverine hypothesis by proposing that Austroasiatic had ultimately expanded into Indochina from the Lingnan area of southern China, with the subsequent Mekong riverine dispersal taking place after the initial arrival of Neolithic farmers from southern China.

Sidwell (2015c) tentatively suggests that Austroasiatic may have begun to split up 5,000 years B.P. during the Neolithic transition era of mainland Southeast Asia, with all the major branches of Austroasiatic formed by 4,000 B.P. Austroasiatic would have had two possible dispersal routes from the western periphery of the Pearl River watershed of Lingnan, which would have been either a coastal route down the coast of Vietnam, or downstream through the Mekong River via Yunnan.^[18] Both the reconstructed lexicon of Proto-Austroasiatic and the archaeological record clearly show that early Austroasiatic speakers around 4,000 B.P. cultivated rice and millet, kept livestock such as dogs, pigs, and chickens, and thrived mostly in estuarine rather than coastal environments.^[18]

At 4,500 B.P., this "Neolithic package" suddenly arrived in Indochina from the Lingnan area without cereal grains and displaced the earlier pre-Neolithic hunter-gatherer cultures, with grain husks found in northern Indochina by 4,100 B.P. and in southern Indochina by 3,800 B.P.^[18] However, Sidwell (2015c) found that iron is not reconstructable in Proto-Austroasiatic, since each Austroasiatic branch has different terms for iron that had been borrowed relatively lately from Tai, Chinese, Tibetan, Malay, and other languages.

During the Iron Age about 2,500 B.P., relatively young Austroasiatic branches in Indochina such as Vietic, Katuic, Pearic, and Khmer were formed, while the more internally diverse Bahnaric branch (dating to about 3,000 B.P.) underwent more extensive internal diversification.^[18] By the Iron Age, all of the Austroasiatic branches were more or less in their present-day locations, with most of the diversification within Austroasiatic taking place during the Iron Age.^[18]

Paul Sidwell (2018)^[19] considers the Austroasiatic language family to have rapidly diversified around 4,000 years B.P. during the arrival of rice agriculture in Indochina, but notes that the origin of Proto-Austroasiatic itself is older than that date. The lexicon of Proto-Austroasiatic can be divided into an early and late stratum. The early stratum consists of basic lexicon including body parts, animal names, natural features, and pronouns, while the names of cultural items (agriculture terms and words for cultural artifacts, which are reconstructible in Proto-Austroasiatic) form part of the later stratum.

Roger Blench (2017)^[20] suggests that vocabulary related to aquatic subsistence strategies (such as boats, waterways, river fauna, and fish capture techniques) can be reconstructed for Proto-Austroasiatic. Blench (2017) finds widespread Austroasiatic roots for 'river, valley', 'boat', 'fish', 'catfish sp.', 'eel', 'prawn', 'shrimp' (Central Austroasiatic), 'crab', 'tortoise', 'turtle', 'otter', 'crocodile', 'heron, fishing bird', and 'fish trap'. Archaeological evidence for the presence of agriculture in northern Indochina (northern Vietnam, Laos, and other nearby areas) dates back to only about 4,000 years ago (2,000 BC), with agriculture ultimately being introduced from further up to the north in the Yangtze valley where it has been dated to 6,000 B.P.^[20]

Sidwell (2022)^[5][21] proposes that the locus of Proto-Austroasiatic was in the Red River Delta area about 4,000-4,500 years before present, instead of the Middle Mekong as he had previously proposed. Austroasiatic dispersed coastal maritime routes and also upstream through river valleys. Khmuic, Palaungic, and Khasic resulted from a westward dispersal that ultimately came from the Red River valley. Based on their current distributions, about half of all Austroasiatic branches (including Nicobaric and Munda) can be traced to coastal maritime dispersals.

Hence, this points to a relatively late riverine dispersal of Austroasiatic as compared to Sino-Tibetan, whose speakers had a distinct non-riverine culture. In addition to living an aquatic-based lifestyle, early Austroasiatic speakers would have also had access to livestock, crops, and newer types of watercraft. As early Austroasiatic speakers dispersed rapidly via waterways, they would have encountered speakers of older language families who were already settled in the area, such as Sino-Tibetan.^[20]

Sidwell (2018)^[22] (quoted in Sidwell 2021^[23]) gives a more nested classification of Austroasiatic branches as suggested by his computational phylogenetic analysis of Austroasiatic languages using a 200-word list. Many of the tentative groupings are likely linkages. Pakanic and Shompen were not included.

Roger Blench (2009)^[24] also proposes that there might have been other primary branches of Austroasiatic that are now extinct, based on substrate evidence in modern-day languages.

• Pre-Chamic languages (the languages of coastal Vietnam before the Chamic migrations). Chamic has various Austroasiatic loanwords that cannot be clearly traced to existing Austroasiatic branches (Sidwell 2006, 2007).^[25][26] Larish (1999)^[27] also notes that Moklenic languages contain many Austroasiatic loanwords, some of which are similar to the ones found in Chamic.
• Acehnese substratum (Sidwell 2006).^[25] Acehnese has many basic words that are of Austroasiatic origin, suggesting that either Austronesian speakers have absorbed earlier Austroasiatic residents in northern Sumatra, or that words might have been borrowed from Austroasiatic languages in southern Vietnam – or perhaps a combination of both. Sidwell (2006) argues that Acehnese and Chamic had often borrowed Austroasiatic words independently of each other, while some Austroasiatic words can be traced back to Proto-Aceh-Chamic. Sidwell (2006) accepts that Acehnese and Chamic are related, but that they had separated from each other before Chamic had borrowed most of its Austroasiatic lexicon.
• Bornean substrate languages (Blench 2010).^[28] Blench cites Austroasiatic-origin words in modern-day Bornean branches such as Land Dayak (Bidayuh, Dayak Bakatiq, etc.), Dusunic (Central Dusun, Visayan, etc.), Kayan, and Kenyah, noting especially resemblances with Aslian. As further evidence for his proposal, Blench also cites ethnographic evidence such as musical instruments in Borneo shared in common with Austroasiatic-speaking groups in mainland Southeast Asia. Adelaar (1995)^[29] has also noticed phonological and lexical similarities between Land Dayak and Aslian. Kaufman (2018) presents dozens of lexical comparisons showing similarities between various Bornean and Austroasiatic languages.^[30]
• Lepcha substratum ("Rongic").^[31] Many words of Austroasiatic origin have been noticed in Lepcha, suggesting a Sino-Tibetan superstrate laid over an Austroasiatic substrate. Blench (2013) calls this branch "Rongic" based on the Lepcha autonym Róng.

Other languages with proposed Austroasiatic substrata are:

• Jiamao, based on evidence from the register system of Jiamao, a Hlai language (Thurgood 1992).^[32] Jiamao is known for its highly aberrant vocabulary in relation to other Hlai languages.
• Kerinci: van Reijn (1974)^[33] notes that Kerinci, a Malayic language of central Sumatra, shares many phonological similarities with Austroasiatic languages, such as sesquisyllabic word structure and vowel inventory.

John Peterson (2017)^[34] suggests that "pre-Munda" (early languages related to Proto-Munda) languages may have once dominated the eastern Indo-Gangetic Plain, and were then absorbed by Indo-Aryan languages at an early date as Indo-Aryan spread east. Peterson notes that eastern Indo-Aryan languages display many morphosyntactic features similar to those of Munda languages, while western Indo-Aryan languages do not.

Other than Latin-based alphabets, many Austroasiatic languages are written with the Khmer, Thai, Lao, and Burmese alphabets. Vietnamese divergently had an indigenous script based on Chinese logographic writing. This has since been supplanted by the Latin alphabet in the 20th century. The following are examples of past-used alphabets or current alphabets of Austroasiatic languages.

• Chữ Nôm (Used until got replaced by Chữ Quốc ngữ (Vietnamese Latin Alphabet) in the 20th century)^[35]
• Khmer alphabet^[36]
• Khom script (used for a short period in the early 20th century for indigenous languages in Laos)
• Old Mon script
• Mon script
• Pahawh Hmong was once used to write Khmu, under the name "Pahawh Khmu"
• Tai Le (Palaung, Blang)
• Tai Tham (Blang)
• Ol Chiki alphabet (Santali alphabet)^[37]
• Mundari Bani (Mundari alphabet)
• Warang Citi (Ho alphabet)^[38]
• Ol Onal (Bhumij alphabet)
• Sorang Sompeng alphabet (Sora alphabet)^[39]

Austroasiatic is an integral part of the controversial Austric hypothesis, which also includes the Austronesian languages, and in some proposals also the Kra–Dai languages and the Hmong–Mien languages.^[40]

Several lexical resemblances are found between the Hmong-Mien and Austroasiatic language families (Ratliff 2010), some of which had earlier been proposed by Haudricourt (1951). This could imply a relation or early language contact along the Yangtze.^[41]

According to Cai (et al. 2011), Hmong–Mien people are genetically related to Austroasiatic speakers, and their languages were heavily influenced by Sino-Tibetan, especially Tibeto-Burman languages.^[42]

It is suggested that the Austroasiatic languages have some influence on Indo-Aryan languages including Sanskrit and middle Indo-Aryan languages. Indian linguist Suniti Kumar Chatterji pointed that a specific number of substantives in languages such as Hindi, Punjabi and Bengali were borrowed from Munda languages. Additionally, French linguist Jean Przyluski suggested a similarity between the tales from the Austroasiatic realm and the Indian mythological stories of Matsyagandha (Satyavati from Mahabharata) and the Nāgas.^[43]

Mitsuru Sakitani suggests that Haplogroup O1b1, which is common in Austroasiatic people and some other ethnic groups in southern China, and haplogroup O1b2, which is common in today's Japanese and Koreans, are the carriers of early rice agriculture from southern China.^[44] Another study suggests that the haplogroup O1b1 is the major Austroasiatic paternal lineage and O1b2 the "para-Austroasiatic" lineage of the Koreans and Yayoi people.^[45]

A full genomic study by Lipson et al. (2018) identified a characteristic lineage that can be associated with the spread of Austroasiatic languages in Southeast Asia and which can be traced back to remains of Neolithic farmers from Mán Bạc (c. 2000 BCE) in the Red River Delta in northern Vietnam, and to closely related Ban Chiang and Vat Komnou remains in Thailand and Cambodia respectively. This Austroasiatic lineage can be modeled as a sister group of the Austronesian peoples with significant admixture (ca. 30%) from a deeply diverging eastern Eurasian source (modeled by the authors as sharing some genetic drift with the Onge, a modern Andamanese hunter-gatherer group) and which is ancestral to modern Austroasiatic-speaking groups of Southeast Asia such as the Mlabri and the Nicobarese, and partially to the Austroasiatic Munda-speaking groups of South Asia (e.g. the Juang). Significant levels of Austroasiatic ancestry were also found in Austronesian-speaking groups of Sumatra, Java, and Borneo.^{[46][note 3]}

Liu et al. (2020) models present Austroasiatic groups from Mainland Southeast Asia as an admixture of Hoabinhian hunter-gatherers and ancestral East Asians associated with the Neolithic farming expansion. Austroasiatic groups cluster with each other except for Kinh Vietnamese and Muong, who share more drift with Tai-Kadai and Hmong-Mien groups.^[48] However, there is evidence of local Austroasiatic input in the Kinh Vietnamese genome.^[49][50] Austroasiatic groups from Southern China, such as the Wa and Blang in Yunnan, predominantly carry the same Mainland Southeast Asian Neolithic farmer ancestry but with additional geneflow from northern and southern East Asian lineages, indicating Tibeto-Burman and Kra-Dai influence respectively.^[51]

Huang et al. (2020) suggests a Southwestern Chinese origin for the 'core Austroasiatic' population, who derive most of their ancestry from Mekong Neolithic (58.0%–75.2%) instead of Late Neolithic Fujian, which is more common for the 'core Austronesian' population. Austroasiatic-related ancestry is widespread in Mainland Southeast Asians and Hmong-Mien groups from Southern China but for the latter, there is evidence of Kra-Dai admixture, which increases in groups that live further east. This admixture is also present in Mainland Southeast Asians. Yangshao culture-related populations, who contributed to the ancestries of present Sino-Tibetan populations, likewise derive their southern East Asian ancestry from Mekong Neolithic (32.2 ± 5.9%).^[52][53] Using Cambodians as proxies for the ancestral Austroasiatic population, they can also be modeled as a mixture of Dai-related groups and groups that are ancestral to all East Asians. The ancestors of Dais themselves can be modeled as a mixture of North Indian-related (6%) and Naxi/Miao-related groups (94%).^[54][55]

According to Kim et al. (2020), Mán Bạc populations constitute the basal ancestry for most populations from Eastern Siberia and Eastern Asia, including Korea, Japan, China and Austroasiatic-speaking groups from Southeast Asia. Populations carrying both Mán Bạc and Devil's Gate genomes admixed throughout these regions until the Neolithic period, which is probably accompanied by climate change and barriers.^[56]

According to Mishra et al. (2024), modern Nicobarese have the highest 'ancestral Austroasiatic' ancestry. This genetic component is found in Austroasiatic populations from South Asia and Southeast Asia.^[57] Another study from 2024, Ahlawat et. al., found that the Austroasiatic tribes — Ho, Bathudi, Bhumij and Mahali from the eastern Indian state of Odisha do not exhibit substantial West Eurasian mtDNA unlike the Dravidian-speaking groups from southern India, and cluster closely with the other Austroasiatic populations of South Asia.^[58]

Wang et al. (2025) states that present Austroasiatic groups are genetically similar to ancient Central Yunnan populations, represented by the Late Neolithic Xingyi individual. This individual has a closer genetic relationship with the Northern East Asian Boshan and the Southern East Asian Qihe3 but is distinct from them. They do not exhibit Basal Asian Xingyi ancestry, which is found in ancient Tibetans, suggesting significant demographic replacement. Ancient individuals from Guangxi like Dushan and Baojianshan, however, have higher affinities with the Qihe3 individual from Fujian and cannot be modeled as having Late Neolithic Xingyi-related ancestry. Alternatively, Central Yunnan populations mediated the expansion of proto-Austroasiatic ancestry in Southeast Asia and Northeast India.^[59]

According to Chaubey et al., "Austro-Asiatic speakers in India today are derived from dispersal from Southeast Asia, followed by extensive sex-specific admixture with local Indian populations."^[60] According to Riccio et al., the Munda peoples are likely descended from Austroasiatic migrants from Southeast Asia.^[61]